Re:Uptime 15,364 days – The Computers of Voyager [video]
by hombre_fatal on Friday October 18, 2019 @09:48AM
(#21291308)
Also, of course there are always exceptions to a platitude. I don't think we need to couch every single statement we ever make with "...but there are exceptions, of course!" which is basically what Knuth goes on to belabor.
Re:Uptime 15,364 days – The Computers of Voyager [video]
by aidenn0 on Friday October 18, 2019 @12:40PM
(#21292820)
More like such platitudes are nearly devoid of meaning:
Premature X is bad.
Overusing X is bad.
These are true for most X. If it's not bad, then you didn't do it prematurely or overuse it!
Re:Uptime 15,364 days – The Computers of Voyager [video]
by aidenn0 on Friday October 18, 2019 @12:39PM
(#21292807)
For further context it was justifying using a goto statement to shave 12% of the execution time off of a function. Knuth bringing it up was specifically to acknowledge that he is aware of the principle to stave off arguments. I more often see it used to push back against any changes for speed.
It's not "optimization is root of all evil". The key is "premature optimization". Maybe people gloss over that part, but it is right there.
Yes, Knuth goes into more detail on what he considers premature optimization in the context of programming computers. However the short sentence applies much more broadly in my experience.
For example, "premature optimization" of BOM costs in a hardware project can cost you dearly down the road when it turns out that leaving in some extra flexibility in the design would be mighty useful.